add docs

* feat: support to set an empty context to the Task

* sytle: fix linter issues

.github/security.md

@@ -1,19 +1,27 @@
-CrewAI takes the security of our software products and services seriously, which includes all source code repositories managed through our GitHub organization.
-If you believe you have found a security vulnerability in any CrewAI product or service, please report it to us as described below.
+## CrewAI Security Vulnerability Reporting Policy
 
- ## Reporting a Vulnerability
- Please do not report security vulnerabilities through public GitHub issues.
- To report a vulnerability, please email us at security@crewai.com.
- Please include the requested information listed below so that we can triage your report more quickly
+CrewAI prioritizes the security of our software products, services, and GitHub repositories. To promptly address vulnerabilities, follow these steps for reporting security issues:
 
- - Type of issue (e.g. SQL injection, cross-site scripting, etc.)
- - Full paths of source file(s) related to the manifestation of the issue
- - The location of the affected source code (tag/branch/commit or direct URL)
- - Any special configuration required to reproduce the issue
- - Step-by-step instructions to reproduce the issue (please include screenshots if needed)
- - Proof-of-concept or exploit code (if possible)
- - Impact of the issue, including how an attacker might exploit the issue
+### Reporting Process
+Do **not** report vulnerabilities via public GitHub issues.
 
- Once we have received your report, we will respond to you at the email address you provide. If the issue is confirmed, we will release a patch as soon as possible depending on the complexity of the issue.
+Email all vulnerability reports directly to:  
+**security@crewai.com**
 
- At this time, we are not offering a bug bounty program. Any rewards will be at our discretion.
+### Required Information
+To help us quickly validate and remediate the issue, your report must include:
+
+- **Vulnerability Type:** Clearly state the vulnerability type (e.g., SQL injection, XSS, privilege escalation).
+- **Affected Source Code:** Provide full file paths and direct URLs (branch, tag, or commit).
+- **Reproduction Steps:** Include detailed, step-by-step instructions. Screenshots are recommended.
+- **Special Configuration:** Document any special settings or configurations required to reproduce.
+- **Proof-of-Concept (PoC):** Provide exploit or PoC code (if available).
+- **Impact Assessment:** Clearly explain the severity and potential exploitation scenarios.
+
+### Our Response
+- We will acknowledge receipt of your report promptly via your provided email.
+- Confirmed vulnerabilities will receive priority remediation based on severity.
+- Patches will be released as swiftly as possible following verification.
+
+### Reward Notice
+Currently, we do not offer a bug bounty program. Rewards, if issued, are discretionary.

.github/workflows/linter.yml

@@ -5,12 +5,29 @@ on: [pull_request]
 jobs:
   lint:
     runs-on: ubuntu-latest
+    env:
+      TARGET_BRANCH: ${{ github.event.pull_request.base.ref }}
     steps:
       - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
 
-      - name: Install Requirements
+      - name: Fetch Target Branch
+        run: git fetch origin $TARGET_BRANCH --depth=1
+
+      - name: Install Ruff
+        run: pip install ruff
+
+      - name: Get Changed Python Files
+        id: changed-files
         run: |
-          pip install ruff
+          merge_base=$(git merge-base origin/"$TARGET_BRANCH" HEAD)
+          changed_files=$(git diff --name-only --diff-filter=ACMRTUB "$merge_base" | grep '\.py$' || true)
+          echo "files<<EOF" >> $GITHUB_OUTPUT
+          echo "$changed_files" >> $GITHUB_OUTPUT
+          echo "EOF" >> $GITHUB_OUTPUT
 
-      - name: Run Ruff Linter
-        run: ruff check
+      - name: Run Ruff on Changed Files
+        if: ${{ steps.changed-files.outputs.files != '' }}
+        run: |
+          echo "${{ steps.changed-files.outputs.files }}" | tr " " "\n" | xargs -I{} ruff check "{}"

.ruff.toml

@@ -2,8 +2,3 @@ exclude = [
     "templates",
     "__init__.py",
 ]
-
-[lint]
-select = [
-    "I", # isort rules
-]

docs/concepts/knowledge.mdx

@@ -700,4 +700,11 @@ recent_news = SpaceNewsKnowledgeSource(
     - Configure appropriate embedding models
     - Consider using local embedding providers for faster processing
   </Accordion>
+
+  <Accordion title="One Time Knowledge">
+    - With the typical file structure provided by CrewAI, knowledge sources are embedded every time the kickoff is triggered.
+    - If the knowledge sources are large, this leads to inefficiency and increased latency, as the same data is embedded each time.
+    - To resolve this, directly initialize the knowledge parameter instead of the knowledge_sources parameter.
+    - Link to the issue to get complete idea [Github Issue](https://github.com/crewAIInc/crewAI/issues/2755)
+  </Accordion>  
 </AccordionGroup>

docs/concepts/llms.mdx

@@ -169,19 +169,55 @@ In this section, you'll find detailed examples that help you select, configure,
     ```
   </Accordion>
 
-  <Accordion title="Google">
-    Set the following environment variables in your `.env` file:
+  <Accordion title="Google (Gemini API)">
+    Set your API key in your `.env` file. If you need a key, or need to find an
+    existing key, check [AI Studio](https://aistudio.google.com/apikey).
 
-    ```toml Code
-    # Option 1: Gemini accessed with an API key.
+    ```toml .env
     # https://ai.google.dev/gemini-api/docs/api-key
     GEMINI_API_KEY=<your-api-key>
-
-    # Option 2: Vertex AI IAM credentials for Gemini, Anthropic, and Model Garden.
-    # https://cloud.google.com/vertex-ai/generative-ai/docs/overview
     ```
 
-    Get credentials from your Google Cloud Console and save it to a JSON file with the following code:
+    Example usage in your CrewAI project:
+    ```python Code
+    from crewai import LLM
+
+    llm = LLM(
+        model="gemini/gemini-2.0-flash",
+        temperature=0.7,
+    )
+    ```
+
+    ### Gemini models
+
+    Google offers a range of powerful models optimized for different use cases.
+
+    | Model                          | Context Window | Best For                                                          |
+    |--------------------------------|----------------|-------------------------------------------------------------------|
+    | gemini-2.5-flash-preview-04-17 | 1M tokens      | Adaptive thinking, cost efficiency                                |
+    | gemini-2.5-pro-preview-05-06   | 1M tokens      | Enhanced thinking and reasoning, multimodal understanding, advanced coding, and more |
+    | gemini-2.0-flash               | 1M tokens      | Next generation features, speed, thinking, and realtime streaming |
+    | gemini-2.0-flash-lite          | 1M tokens      | Cost efficiency and low latency                                   |
+    | gemini-1.5-flash               | 1M tokens      | Balanced multimodal model, good for most tasks                    |
+    | gemini-1.5-flash-8B            | 1M tokens      | Fastest, most cost-efficient, good for high-frequency tasks       |
+    | gemini-1.5-pro                 | 2M tokens      | Best performing, wide variety of reasoning tasks including logical reasoning, coding, and creative collaboration |
+
+    The full list of models is available in the [Gemini model docs](https://ai.google.dev/gemini-api/docs/models).
+
+    ### Gemma
+
+    The Gemini API also allows you to use your API key to access [Gemma models](https://ai.google.dev/gemma/docs) hosted on Google infrastructure.
+
+    | Model          | Context Window |
+    |----------------|----------------|
+    | gemma-3-1b-it  | 32k tokens     |
+    | gemma-3-4b-it  | 32k tokens     |
+    | gemma-3-12b-it | 32k tokens     |
+    | gemma-3-27b-it | 128k tokens    |
+
+  </Accordion>
+  <Accordion title="Google (Vertex AI)">
+    Get credentials from your Google Cloud Console and save it to a JSON file, then load it with the following code:
     ```python Code
     import json
 
@@ -205,14 +241,18 @@ In this section, you'll find detailed examples that help you select, configure,
         vertex_credentials=vertex_credentials_json
     )
     ```
+
     Google offers a range of powerful models optimized for different use cases:
 
-    | Model                  | Context Window | Best For                                                          |
-    |-----------------------|----------------|------------------------------------------------------------------|
-    | gemini-2.0-flash-exp  | 1M tokens      | Higher quality at faster speed, multimodal model, good for most tasks |
-    | gemini-1.5-flash      | 1M tokens      | Balanced multimodal model, good for most tasks                    |
-    | gemini-1.5-flash-8B   | 1M tokens      | Fastest, most cost-efficient, good for high-frequency tasks       |
-    | gemini-1.5-pro        | 2M tokens      | Best performing, wide variety of reasoning tasks including logical reasoning, coding, and creative collaboration |
+    | Model                          | Context Window | Best For                                                          |
+    |--------------------------------|----------------|-------------------------------------------------------------------|
+    | gemini-2.5-flash-preview-04-17 | 1M tokens      | Adaptive thinking, cost efficiency                                |
+    | gemini-2.5-pro-preview-05-06   | 1M tokens      | Enhanced thinking and reasoning, multimodal understanding, advanced coding, and more |
+    | gemini-2.0-flash               | 1M tokens      | Next generation features, speed, thinking, and realtime streaming |
+    | gemini-2.0-flash-lite          | 1M tokens      | Cost efficiency and low latency                                   |
+    | gemini-1.5-flash               | 1M tokens      | Balanced multimodal model, good for most tasks                    |
+    | gemini-1.5-flash-8B            | 1M tokens      | Fastest, most cost-efficient, good for high-frequency tasks       |
+    | gemini-1.5-pro                 | 2M tokens      | Best performing, wide variety of reasoning tasks including logical reasoning, coding, and creative collaboration |
   </Accordion>
 
   <Accordion title="Azure">

docs/enterprise/features/agent-repository.mdx

@@ -0,0 +1,70 @@
+---
+title: "Agent Repository"
+description: "Store and retrieve agents for your CrewAI projects"
+---
+
+# Agent Repository
+
+The Agent Repository allows you to store, manage, and reuse agents across your CrewAI projects. This feature streamlines the development process by enabling you to configure agents once and use them in multiple projects.
+
+## How It Works
+
+When you create an agent in the CrewAI interface, it's stored in the Agent Repository. You can then initialize these agents in your code using the `from_repository` parameter.
+
+## Usage
+
+To use an agent from the repository in your CrewAI project, initialize it with the following code:
+
+```python
+from crewai import Agent
+
+# Initialize the agent with its role
+agent = Agent(from_repository="python-job-researcher")
+```
+
+### Creating a Crew with Repository Agents
+
+```python
+from crewai import Agent, Crew, Task
+
+agent = Agent(from_repository="python-job-researcher")
+
+job_search_task = Task(
+    description="Search for recent Python developer job listings online",
+    expected_output="Markdown list of 5 recent Python developer jobs with details.",
+    agent=agent,
+)
+
+crew = Crew(agents=[agent], tasks=[job_search_task], verbose=True)
+
+result = crew.kickoff()
+print(result)
+```
+
+## Important Notes
+
+- The `from_repository` value must match the agent's role in a URL-safe format.
+- If you change an agent's role after creation, you must update the `from_repository` value in your code accordingly, or you won't be able to find the agent anymore.
+- Make sure you have permission to use the agent as mentioned in the key points.
+
+## Agent Configuration
+
+When configuring an agent in the repository, you can specify:
+
+1. **Role** - The agent's primary function (e.g., "Python Job Researcher")
+2. **Goal** - What the agent aims to achieve (e.g., "Find Python developer job opportunities")
+3. **Backstory** - Context for the agent's behavior
+4. **Tools** - Available capabilities for the agent to use when performing tasks
+5. **Visibility Controls** - Who can access and use the agent
+
+## Managing Agents
+
+The Agent Repository interface provides functionality to:
+
+- View all available agents
+- Add new agents
+- Edit existing agents
+- Delete agents
+- View agent details including usage examples
+
+By leveraging the Agent Repository, you can build more modular and reusable AI workflows while maintaining a central location for managing your agents.

docs/how-to/arize-phoenix-observability.mdx

@@ -68,7 +68,13 @@ We'll create a CrewAI application where two agents collaborate to research and w
 ```python
 from crewai import Agent, Crew, Process, Task
 from crewai_tools import SerperDevTool
+from openinference.instrumentation.crewai import CrewAIInstrumentor
+from phoenix.otel import register
 
+# setup monitoring for your crew
+tracer_provider = register(
+    endpoint="http://localhost:6006/v1/traces")
+CrewAIInstrumentor().instrument(skip_dep_check=True, tracer_provider=tracer_provider)
 search_tool = SerperDevTool()
 
 # Define your agents with roles and goals

src/crewai/agent.py

@@ -20,6 +20,7 @@ from crewai.tools.agent_tools.agent_tools import AgentTools
 from crewai.utilities import Converter, Prompts
 from crewai.utilities.agent_utils import (
     get_tool_names,
+    load_agent_from_repository,
     parse_tools,
     render_text_description_and_args,
 )
@@ -134,6 +135,16 @@ class Agent(BaseAgent):
         default=None,
         description="Knowledge search query for the agent dynamically generated by the agent.",
     )
+    from_repository: Optional[str] = Field(
+        default=None,
+        description="The Agent's role to be used from your repository.",
+    )
+
+    @model_validator(mode="before")
+    def validate_from_repository(cls, v):
+        if v is not None and (from_repository := v.get("from_repository")):
+            return load_agent_from_repository(from_repository) | v
+        return v
 
     @model_validator(mode="after")
     def post_init_setup(self):

src/crewai/cli/authentication/token.py

@@ -5,5 +5,5 @@ def get_auth_token() -> str:
     """Get the authentication token."""
     access_token = TokenManager().get_token()
     if not access_token:
-        raise Exception()
+        raise Exception("No token found, make sure you are logged in")
     return access_token

src/crewai/cli/constants.py

@@ -13,7 +13,7 @@ ENV_VARS = {
     ],
     "gemini": [
         {
-            "prompt": "Enter your GEMINI API key (press Enter to skip)",
+            "prompt": "Enter your GEMINI API key from https://ai.dev/apikey (press Enter to skip)",
             "key_name": "GEMINI_API_KEY",
         }
     ],

src/crewai/cli/plus_api.py

@@ -14,6 +14,7 @@ class PlusAPI:
 
     TOOLS_RESOURCE = "/crewai_plus/api/v1/tools"
     CREWS_RESOURCE = "/crewai_plus/api/v1/crews"
+    AGENTS_RESOURCE = "/crewai_plus/api/v1/agents"
 
     def __init__(self, api_key: str) -> None:
         self.api_key = api_key
@@ -37,6 +38,9 @@ class PlusAPI:
     def get_tool(self, handle: str):
         return self._make_request("GET", f"{self.TOOLS_RESOURCE}/{handle}")
 
+    def get_agent(self, handle: str):
+        return self._make_request("GET", f"{self.AGENTS_RESOURCE}/{handle}")
+
     def publish_tool(
         self,
         handle: str,

src/crewai/crew.py

@@ -52,7 +52,7 @@ from crewai.tools.agent_tools.agent_tools import AgentTools
 from crewai.tools.base_tool import BaseTool, Tool
 from crewai.types.usage_metrics import UsageMetrics
 from crewai.utilities import I18N, FileHandler, Logger, RPMController
-from crewai.utilities.constants import TRAINING_DATA_FILE
+from crewai.utilities.constants import NOT_SPECIFIED, TRAINING_DATA_FILE
 from crewai.utilities.evaluators.crew_evaluator_handler import CrewEvaluator
 from crewai.utilities.evaluators.task_evaluator import TaskEvaluator
 from crewai.utilities.events.crew_events import (
@@ -478,7 +478,7 @@ class Crew(FlowTrackable, BaseModel):
         separated by a synchronous task.
         """
         for i, task in enumerate(self.tasks):
-            if task.async_execution and task.context:
+            if task.async_execution and isinstance(task.context, list):
                 for context_task in task.context:
                     if context_task.async_execution:
                         for j in range(i - 1, -1, -1):
@@ -496,7 +496,7 @@ class Crew(FlowTrackable, BaseModel):
         task_indices = {id(task): i for i, task in enumerate(self.tasks)}
 
         for task in self.tasks:
-            if task.context:
+            if isinstance(task.context, list):
                 for context_task in task.context:
                     if id(context_task) not in task_indices:
                         continue  # Skip context tasks not in the main tasks list
@@ -1034,11 +1034,14 @@ class Crew(FlowTrackable, BaseModel):
                 )
         return cast(List[BaseTool], tools)
 
-    def _get_context(self, task: Task, task_outputs: List[TaskOutput]):
+    def _get_context(self, task: Task, task_outputs: List[TaskOutput]) -> str:
+        if not task.context:
+            return ""
+
         context = (
-            aggregate_raw_outputs_from_tasks(task.context)
-            if task.context
-            else aggregate_raw_outputs_from_task_outputs(task_outputs)
+            aggregate_raw_outputs_from_task_outputs(task_outputs)
+            if task.context is NOT_SPECIFIED
+            else aggregate_raw_outputs_from_tasks(task.context)
         )
         return context
 
@@ -1226,7 +1229,7 @@ class Crew(FlowTrackable, BaseModel):
             task_mapping[task.key] = cloned_task
 
         for cloned_task, original_task in zip(cloned_tasks, self.tasks):
-            if original_task.context:
+            if isinstance(original_task.context, list):
                 cloned_context = [
                     task_mapping[context_task.key]
                     for context_task in original_task.context

src/crewai/llm.py

@@ -5,8 +5,7 @@ import sys
 import threading
 import warnings
 from collections import defaultdict
-from contextlib import contextmanager
-from types import SimpleNamespace
+from contextlib import contextmanager, redirect_stderr, redirect_stdout
 from typing import (
     Any,
     DefaultDict,
@@ -31,7 +30,6 @@ from crewai.utilities.events.llm_events import (
     LLMCallType,
     LLMStreamChunkEvent,
 )
-from crewai.utilities.events.tool_usage_events import ToolExecutionErrorEvent
 
 with warnings.catch_warnings():
     warnings.simplefilter("ignore", UserWarning)
@@ -45,6 +43,9 @@ with warnings.catch_warnings():
     from litellm.utils import supports_response_schema
 
 
+import io
+from typing import TextIO
+
 from crewai.llms.base_llm import BaseLLM
 from crewai.utilities.events import crewai_event_bus
 from crewai.utilities.exceptions.context_window_exceeding_exception import (
@@ -54,12 +55,17 @@ from crewai.utilities.exceptions.context_window_exceeding_exception import (
 load_dotenv()
 
 
-class FilteredStream:
-    def __init__(self, original_stream):
+class FilteredStream(io.TextIOBase):
+    _lock = None
+
+    def __init__(self, original_stream: TextIO):
         self._original_stream = original_stream
         self._lock = threading.Lock()
 
-    def write(self, s) -> int:
+    def write(self, s: str) -> int:
+        if not self._lock:
+            self._lock = threading.Lock()
+
         with self._lock:
             # Filter out extraneous messages from LiteLLM
             if (
@@ -214,15 +220,11 @@ def suppress_warnings():
         )
 
         # Redirect stdout and stderr
-        old_stdout = sys.stdout
-        old_stderr = sys.stderr
-        sys.stdout = FilteredStream(old_stdout)
-        sys.stderr = FilteredStream(old_stderr)
-        try:
+        with (
+            redirect_stdout(FilteredStream(sys.stdout)),
+            redirect_stderr(FilteredStream(sys.stderr)),
+        ):
             yield
-        finally:
-            sys.stdout = old_stdout
-            sys.stderr = old_stderr
 
 
 class Delta(TypedDict):

src/crewai/task.py

@@ -2,7 +2,6 @@ import datetime
 import inspect
 import json
 import logging
-import re
 import threading
 import uuid
 from concurrent.futures import Future
@@ -41,6 +40,7 @@ from crewai.tasks.output_format import OutputFormat
 from crewai.tasks.task_output import TaskOutput
 from crewai.tools.base_tool import BaseTool
 from crewai.utilities.config import process_config
+from crewai.utilities.constants import NOT_SPECIFIED
 from crewai.utilities.converter import Converter, convert_to_model
 from crewai.utilities.events import (
     TaskCompletedEvent,
@@ -97,7 +97,7 @@ class Task(BaseModel):
     )
     context: Optional[List["Task"]] = Field(
         description="Other tasks that will have their output used as context for this task.",
-        default=None,
+        default=NOT_SPECIFIED,
     )
     async_execution: Optional[bool] = Field(
         description="Whether the task should be executed asynchronously or not.",
@@ -643,7 +643,7 @@ class Task(BaseModel):
 
         cloned_context = (
             [task_mapping[context_task.key] for context_task in self.context]
-            if self.context
+            if isinstance(self.context, list)
             else None
         )
 

src/crewai/telemetry/telemetry.py

@@ -10,6 +10,18 @@ from contextlib import contextmanager
 from importlib.metadata import version
 from typing import TYPE_CHECKING, Any, Optional
 
+from opentelemetry import trace
+from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
+    OTLPSpanExporter,
+)
+from opentelemetry.sdk.resources import SERVICE_NAME, Resource
+from opentelemetry.sdk.trace import TracerProvider
+from opentelemetry.sdk.trace.export import (
+    BatchSpanProcessor,
+    SpanExportResult,
+)
+from opentelemetry.trace import Span, Status, StatusCode
+
 from crewai.telemetry.constants import (
     CREWAI_TELEMETRY_BASE_URL,
     CREWAI_TELEMETRY_SERVICE_NAME,
@@ -25,18 +37,6 @@ def suppress_warnings():
         yield
 
 
-from opentelemetry import trace  # noqa: E402
-from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
-    OTLPSpanExporter,  # noqa: E402
-)
-from opentelemetry.sdk.resources import SERVICE_NAME, Resource  # noqa: E402
-from opentelemetry.sdk.trace import TracerProvider  # noqa: E402
-from opentelemetry.sdk.trace.export import (  # noqa: E402
-    BatchSpanProcessor,
-    SpanExportResult,
-)
-from opentelemetry.trace import Span, Status, StatusCode  # noqa: E402
-
 if TYPE_CHECKING:
     from crewai.crew import Crew
     from crewai.task import Task
@@ -232,7 +232,7 @@ class Telemetry:
                                 "agent_key": task.agent.key if task.agent else None,
                                 "context": (
                                     [task.description for task in task.context]
-                                    if task.context
+                                    if isinstance(task.context, list)
                                     else None
                                 ),
                                 "tools_names": [
@@ -748,7 +748,7 @@ class Telemetry:
                             "agent_key": task.agent.key if task.agent else None,
                             "context": (
                                 [task.description for task in task.context]
-                                if task.context
+                                if isinstance(task.context, list)
                                 else None
                             ),
                             "tools_names": [

src/crewai/utilities/agent_utils.py

@@ -16,6 +16,7 @@ from crewai.tools.base_tool import BaseTool
 from crewai.tools.structured_tool import CrewStructuredTool
 from crewai.tools.tool_types import ToolResult
 from crewai.utilities import I18N, Printer
+from crewai.utilities.errors import AgentRepositoryError
 from crewai.utilities.exceptions.context_window_exceeding_exception import (
     LLMContextLengthExceededException,
 )
@@ -428,3 +429,36 @@ def show_agent_logs(
             printer.print(
                 content=f"\033[95m## Final Answer:\033[00m \033[92m\n{formatted_answer.output}\033[00m\n\n"
             )
+
+
+def load_agent_from_repository(from_repository: str) -> Dict[str, Any]:
+    attributes: Dict[str, Any] = {}
+    if from_repository:
+        import importlib
+
+        from crewai.cli.authentication.token import get_auth_token
+        from crewai.cli.plus_api import PlusAPI
+
+        client = PlusAPI(api_key=get_auth_token())
+        response = client.get_agent(from_repository)
+        if response.status_code != 200:
+            raise AgentRepositoryError(
+                f"Agent {from_repository} could not be loaded: {response.text}"
+            )
+
+        agent = response.json()
+        for key, value in agent.items():
+            if key == "tools":
+                attributes[key] = []
+                for tool_name in value:
+                    try:
+                        module = importlib.import_module("crewai_tools")
+                        tool_class = getattr(module, tool_name)
+                        attributes[key].append(tool_class())
+                    except Exception as e:
+                        raise AgentRepositoryError(
+                            f"Tool {tool_name} could not be loaded: {e}"
+                        ) from e
+            else:
+                attributes[key] = value
+    return attributes

src/crewai/utilities/constants.py

@@ -5,3 +5,14 @@ KNOWLEDGE_DIRECTORY = "knowledge"
 MAX_LLM_RETRY = 3
 MAX_FILE_NAME_LENGTH = 255
 EMITTER_COLOR = "bold_blue"
+
+
+class _NotSpecified:
+    def __repr__(self):
+        return "NOT_SPECIFIED"
+
+
+# Sentinel value used to detect when no value has been explicitly provided.
+# Unlike `None`, which might be a valid value from the user, `NOT_SPECIFIED` allows
+# us to distinguish between "not passed at all" and "explicitly passed None" or "[]".
+NOT_SPECIFIED = _NotSpecified()

src/crewai/utilities/errors.py

@@ -1,4 +1,5 @@
 """Error message definitions for CrewAI database operations."""
+
 from typing import Optional
 
 
@@ -37,3 +38,9 @@ class DatabaseError:
             The formatted error message
         """
         return template.format(str(error))
+
+
+class AgentRepositoryError(Exception):
+    """Exception raised when an agent repository is not found."""
+
+    ...

src/crewai/utilities/formatter.py

@@ -1,6 +1,6 @@
-import re
 from typing import TYPE_CHECKING, List
 
+
 if TYPE_CHECKING:
     from crewai.task import Task
     from crewai.tasks.task_output import TaskOutput
@@ -17,6 +17,11 @@ def aggregate_raw_outputs_from_task_outputs(task_outputs: List["TaskOutput"]) ->
 
 def aggregate_raw_outputs_from_tasks(tasks: List["Task"]) -> str:
     """Generate string context from the tasks."""
-    task_outputs = [task.output for task in tasks if task.output is not None]
+
+    task_outputs = (
+        [task.output for task in tasks if task.output is not None]
+        if isinstance(tasks, list)
+        else []
+    )
 
     return aggregate_raw_outputs_from_task_outputs(task_outputs)

tests/agent_test.py

@@ -2,7 +2,7 @@
 
 import os
 from unittest import mock
-from unittest.mock import patch
+from unittest.mock import MagicMock, patch
 
 import pytest
 
@@ -18,6 +18,7 @@ from crewai.tools import tool
 from crewai.tools.tool_calling import InstructorToolCalling
 from crewai.tools.tool_usage import ToolUsage
 from crewai.utilities import RPMController
+from crewai.utilities.errors import AgentRepositoryError
 from crewai.utilities.events import crewai_event_bus
 from crewai.utilities.events.tool_usage_events import ToolUsageFinishedEvent
 
@@ -308,9 +309,7 @@ def test_cache_hitting():
     def handle_tool_end(source, event):
         received_events.append(event)
 
-    with (
-        patch.object(CacheHandler, "read") as read,
-    ):
+    with (patch.object(CacheHandler, "read") as read,):
         read.return_value = "0"
         task = Task(
             description="What is 2 times 6? Ignore correctness and just return the result of the multiplication tool, you must use the tool.",
@@ -1040,7 +1039,7 @@ def test_agent_human_input():
             CrewAgentExecutor,
             "_invoke_loop",
             return_value=AgentFinish(output="Hello", thought="", text=""),
-        ) as mock_invoke_loop,
+        ),
     ):
         # Execute the task
         output = agent.execute_task(task)
@@ -2025,3 +2024,86 @@ def test_get_knowledge_search_query():
                 },
             ]
         )
+
+
+@pytest.fixture
+def mock_get_auth_token():
+    with patch(
+        "crewai.cli.authentication.token.get_auth_token", return_value="test_token"
+    ):
+        yield
+
+
+@patch("crewai.cli.plus_api.PlusAPI.get_agent")
+def test_agent_from_repository(mock_get_agent, mock_get_auth_token):
+    from crewai_tools import SerperDevTool
+
+    mock_get_response = MagicMock()
+    mock_get_response.status_code = 200
+    mock_get_response.json.return_value = {
+        "role": "test role",
+        "goal": "test goal",
+        "backstory": "test backstory",
+        "tools": ["SerperDevTool"],
+    }
+    mock_get_agent.return_value = mock_get_response
+    agent = Agent(from_repository="test_agent")
+
+    assert agent.role == "test role"
+    assert agent.goal == "test goal"
+    assert agent.backstory == "test backstory"
+    assert len(agent.tools) == 1
+    assert isinstance(agent.tools[0], SerperDevTool)
+
+
+@patch("crewai.cli.plus_api.PlusAPI.get_agent")
+def test_agent_from_repository_override_attributes(mock_get_agent, mock_get_auth_token):
+    from crewai_tools import SerperDevTool
+
+    mock_get_response = MagicMock()
+    mock_get_response.status_code = 200
+    mock_get_response.json.return_value = {
+        "role": "test role",
+        "goal": "test goal",
+        "backstory": "test backstory",
+        "tools": ["SerperDevTool"],
+    }
+    mock_get_agent.return_value = mock_get_response
+    agent = Agent(from_repository="test_agent", role="Custom Role")
+
+    assert agent.role == "Custom Role"
+    assert agent.goal == "test goal"
+    assert agent.backstory == "test backstory"
+    assert len(agent.tools) == 1
+    assert isinstance(agent.tools[0], SerperDevTool)
+
+
+@patch("crewai.cli.plus_api.PlusAPI.get_agent")
+def test_agent_from_repository_with_invalid_tools(mock_get_agent, mock_get_auth_token):
+    mock_get_response = MagicMock()
+    mock_get_response.status_code = 200
+    mock_get_response.json.return_value = {
+        "role": "test role",
+        "goal": "test goal",
+        "backstory": "test backstory",
+        "tools": ["DoesNotExist"],
+    }
+    mock_get_agent.return_value = mock_get_response
+    with pytest.raises(
+        AgentRepositoryError,
+        match="Tool DoesNotExist could not be loaded: module 'crewai_tools' has no attribute 'DoesNotExist'",
+    ):
+        Agent(from_repository="test_agent")
+
+
+@patch("crewai.cli.plus_api.PlusAPI.get_agent")
+def test_agent_from_repository_agent_not_found(mock_get_agent, mock_get_auth_token):
+    mock_get_response = MagicMock()
+    mock_get_response.status_code = 404
+    mock_get_response.text = "Agent not found"
+    mock_get_agent.return_value = mock_get_response
+    with pytest.raises(
+        AgentRepositoryError,
+        match="Agent NOT_FOUND could not be loaded: Agent not found",
+    ):
+        Agent(from_repository="NOT_FOUND")

tests/crew_test.py

@@ -2,22 +2,18 @@
 
 import hashlib
 import json
-import os
-import tempfile
 from concurrent.futures import Future
 from unittest import mock
-from unittest.mock import MagicMock, patch
+from unittest.mock import ANY, MagicMock, patch
 
 import pydantic_core
 import pytest
 
 from crewai.agent import Agent
 from crewai.agents import CacheHandler
-from crewai.agents.cache import CacheHandler
-from crewai.agents.crew_agent_executor import CrewAgentExecutor
 from crewai.crew import Crew
 from crewai.crews.crew_output import CrewOutput
-from crewai.flow import Flow, listen, start
+from crewai.flow import Flow, start
 from crewai.knowledge.source.string_knowledge_source import StringKnowledgeSource
 from crewai.llm import LLM
 from crewai.memory.contextual.contextual_memory import ContextualMemory
@@ -3141,6 +3137,30 @@ def test_replay_with_context():
         assert crew.tasks[1].context[0].output.raw == "context raw output"
 
 
+def test_replay_with_context_set_to_nullable():
+    agent = Agent(role="test_agent", backstory="Test Description", goal="Test Goal")
+    task1 = Task(
+        description="Context Task", expected_output="Say Task Output", agent=agent
+    )
+    task2 = Task(
+        description="Test Task", expected_output="Say Hi", agent=agent, context=[]
+    )
+    task3 = Task(
+        description="Test Task 3", expected_output="Say Hi", agent=agent, context=None
+    )
+
+    crew = Crew(agents=[agent], tasks=[task1, task2, task3], process=Process.sequential)
+    with patch("crewai.task.Task.execute_sync") as mock_execute_task:
+        mock_execute_task.return_value = TaskOutput(
+            description="Test Task Output",
+            raw="test raw output",
+            agent="test_agent",
+        )
+        crew.kickoff()
+
+    mock_execute_task.assert_called_with(agent=ANY, context="", tools=ANY)
+
+
 @pytest.mark.vcr(filter_headers=["authorization"])
 def test_replay_with_invalid_task_id():
     agent = Agent(role="test_agent", backstory="Test Description", goal="Test Goal")